260 likes | 424 Views
Lecture 11. Dynamic Asset Pricing Models - II. Fixing the C-CAPM. The risk-premium puzzle is a big drag on structural models, like the C-CAPM, which are loved by economists. A lot of efforts to salvage them:
E N D
Lecture 11 Dynamic Asset Pricing Models - II
Fixing the C-CAPM The risk-premium puzzle is a big drag on structural models, like the C-CAPM, which are loved by economists. A lot of efforts to salvage them: • (1) Power utility function is too strict, since EIS and γ are unrealistically linked. Epstein-Zin-Weil’s effort. Hansen, Heaton and Li (2007). • (2) Model risk aversion parameter. A time-varying γt?. Constantinides’s (1990) habit formation. Abel (1990), Campbell and Cochrane (1999).
(3) Use different consumption goods: durables and non-durables. Eichenbaum and Hansen (1990), Pakos (2004), Yogo (2004). • (4) Problem with Et[rt+1]. Ex-ante and ex-post are different. Soderlind’s (2006) survey data. Peso problem? Rietz’s (1988) disaster state. Also, add learning to model: Bayesian learning. Latest effort: Cogley and Sargent (2008). Survival bias? Brown, Goetzmann and Ross (1995).
Habit Formation • Constantinides (1990): Average Joe’s utility depends on the difference between current consumption and the “habit”: U(c) = EtΣiβi [(ct+i – xt+i)1-γ]/(1-γ), λ >0. Let the habit depend on lagged consumption (an internal habit): xt+1= λct+i-1 U’(ct+i) = Et+i {βi (ct+i – λct+i-1)-γ+ βi+1 (ct+i+1 – λct+i)-γ(- λ)} = βict+i-1-γ((ct+i/ct+i-1)–λ)-γ– λβi+1ct+i-γEt+i[((ct+i+1/ct+i)–λ)-γ] = βict+i-1-γ{(gt+i–λ)-γ– λβgt+i-γEt+i[(gt+i+1–λ)-γ]} (gt+i=ct+i/ct+i-1) If we assume that gt+i is iid (not a realistic assumption): Et+i[(gt+i+1–λ)-γ] = Et[(gt+1–λ)-γ] = E[(gt–λ)-γ] = Θ (a constant)
Then, recall that mt+1 = U’(ct+1)/ U’(ct) mt+1= β gt-γ{(gt+1–λ)-γ– λβ gt+1-γΘ}/{(gt–λ)-γ– λβ gt-γΘ} Compare this to mt+1 from the standard model: mt+1 = βgt+1-γ • Now, since λ>0, and (gt+1–λ) < 1 is more often than gt+1 < 1. Thus, we do not need a large risk-aversion coefficient to amplify the variation of consumption. Note: If λ =1 and the habit is fixed, the utility function becomes: U(c) = EtΣiβi [(ct+i – x)1-γ]/(1-γ) x: fixed subsistence level (habit). RRA = γ /(1 - x/ct) (if x/ct=0.8, RRA= 5 γ.) • But, risk aversion is now time-varying. This utility function makes the agent very averse to consumption risk. It helps to address the low risk-free rate.
Campbell and Cochrane’s (1999): Take on Abel’s (1990) formulation of habit formation. Let U(c) be • U(c) = EtΣiβi [(Ct+i – Xt+i)1-γ – 1 ]/(1-γ), • where xt is the level of habit. • CC work with “surplus consumption ratio” –their state variable: • St = (Ct – Xt)/Ct • CC define the habit xt as external: Average Joe looks at aggregate consumption (the “Joneses”) to determine his level of happiness: • Sat = [(Cat – Xt)/Cat • Now, mt+1 = β[(Ct+1/Ct)(St+1/St)]-γ • CC assume a st= log(St) follows an heteroscedastic AR(1) process: • sat = (1-φ) s + φ sat-1 + λ(sat) (cat – cat-1 – g) • CC assume Δct=log(Ct/Ct-1) ~ iid lognormal (g,σ2).
Now, the risk-free rate: • rf = - ln(β) + γg - γ(1-φ) (st-s) – γ2σ2/2 [1+λ(st)]2 • CC have to choose λ(st) to proceed. They use a simple threshold function –i.e., with two (implicit) surplus states. • CC explain the volatility puzzle and the low interest rate puzzle. But they also need a high risk aversion coefficient to match the equity premium (average risk aversion 80, in low surplus state gets to 100!). • CC find that agents fear stocks primarily because they do badly in recessions (times of low surplus consumption ratios), not because stock returns are correlated with declines in wealth or consumption. • CC conclude that habit formation, or some other device is needed to generate time-varying countercyclical risk premia along with relatively constant risk-free rates.
Separating EIS (ψ) and γ • Epstein and Zin (1991) – Weil (1989): They work with a recursive utility function: Ut = {(1-δ) Ct(1-γ)/θ+ δ (Et[ Ut+1(1-γ) ]1/θ} θ/(1-γ) where θ = (1-γ)/(1-1/ψ). When θ=1, the recursion becomes linear (usual power utility model). • The intertemporal budget constraint is Wt+1 = (1+Rm,t+1) (Wt - Ct) (Wt = Average Joe’s wealth.) • Assuming lognormality and homoscedasticity: rf,t+1 = - ln δ + (θ-1)/2 σm2 - θ /(2ψ2)σc2 + 1/ψ Et[ln(ct+1) - ln(ct)]
Then, the risk premium for asset i is: • E[ri,t+1] – rf,t+1 = θ/ψσic– (1 – θ)σi,m –σc2 /2 • The Epstein-Zin-Weil model nests the static CAPM (θ=0) and the C-CAPM with power utility ((θ=1). • Campbell (1993) introduces a log-linearized budget constraint and after a lot of substitutions, we get: • E[ri,t+1] – rf,t+1 = γσim + (1 – γ)σih– σi2 /2, • where σih is the covariance of asset i with “news” about future returns on the market –see CLM. • Static CAPM? - γ = 1 (no realistic) • - σih= 0 (no realistic) • - Rm,t follows a univariate process, then future returns are perfectly correlated with current return. (Maybe?).
• Findings: Equity premium estimates using this model are lower γ. But, with an unrealistic consumption volatility.
More General Utility Functions • Introduce nonseparabilities: consumption and some other good. • Usually done using Cobb-Douglas utility functions. Then, we have an easy to work marginal utility function: u’(Ct,Xt) = Ct-γ1 Xt-γ2 where Xt represents some other good. • Now, we have an easy to work Euler’s equation: 1= Et[β(Ct+1/Ct)-γ1(Xt+1/Xt)–γ2 (1+Rt+1]. Assuming joint lognormality and homoscedasticity: Et[ri,t+1] = μi + γ1 Et[Δct+1] + γ2Et[Δxt+1] ,
Other goods: leisure -Eichenbaum, Hansen and Singleton (1988) government spending – Aschauer (1985) stock of durable goods – Startz (1989) Findings: None of the Xt variables significantly improves the CAPM. • Surprised? Not really. The proposed Xt variables do not have enough variability to explain the variability of excess returns.
Ex-ante ≠ Ex-post: Survey Evidence • C-CAPM: E[rt] – rf,t = Cov(rt, Δct) γ = = Corr(rt, Δct) σcσmγ • Fitting U.S. ex-post data (1952–2005) creates the “EP puzzle:” E[rt] – rf,t = .06 = Corr(rt, Δct) σcσmγ = 0.17 (or 0.33) x 0.14 x 0.01 x γ => Requires risk aversion, γ, to be very high: γ≥ 43! • Soderling (2006) tests the C-CAPM using ex-ante data: - For E[rt] – rf,t, survey data (Livingston survery (1952-2005) - For σc, survey data (SPF: survey of professional forecasters) - For σm, no survey data. Expectations extracted from S&P options. - For Corr(rt, Δct), no expectations data.
This approach avoids the usual “joint test:” C-CAPM and RE. • Data: • Livingston survey: survey of economic experts (1952–2005). Before 1990: S&P Industrials; After 1990, S&P 500 Composite. • SPF: survey of economic experts about probabilities of GDP growth rates –assume results carry over to consumption growth. • Lots of S&P 500 options (daily data 1985-2005). • Example: Excess returns survey
• E[rt] – rf,t≈ 3% - 3.5% (about half the historical mean excess return) • Volatile market, hard to “learn.” Test for zero average error. • Average forecast error = 3%, SD of forecast error = 16%, T=100 • => t = 3/(16/1001/2) = 1.9
• Ex-ante estimates: • - E[rt] – rf,t≈ 0.5 x 0.06 = .03 • - σc≈ 0.6 x 0.01 = .006 • - σm, ≈ 1.15 x 0.14 = .161 • - Corr(rt, Δct) = ? . • E[rt] – rf,t = .03 = Corr(rt, Δct) σcσmγ = 0.17 (or 0.33) x 0.161 x 0.006x γ • => Now, γ needs to be 0.7 of γ in ex-post data (30 instead of 43) • Better for C-CAPM, but still a very high γ
Learning • Cogley and Sargent (2008): Follow Friedman and Schwartz (1963), where the U.S. 1930 depression created a pessimist mood, seriously affected estimates of expected returns. (The “depression generation.”) • Siegel (1992) presents supporting evidence: 1802-1925 Equity premium was 2.0%; 1926-1990 Equity premium was 5.9%. • CS present a learning model to explain the equity premium. • Assumptions: • Consumption is driven by a two-state Markov process. • Representative agent is a Bayesian learner. • Initial beliefs from the 1930s are very pessimistic. • Learning is slow. • Asset pricing is distorted by these beliefs for a long time.
Related work: Cecchetti, Lam, Mark (2000). • Equity premium due to distorted beliefs. CLM (2000) have no learning. Agents are naturally and permanently pessimistic. • Under Euler’s equation we have: • pt = Ets[mt+1 xt+1] • If Etsis the expectation implied by the true transition probabilities F, the agent has rational expectations. Call this Eta. The equity premium will be small in that case. • CLM (2000) show that Ets ≠Eta explain equity premium. But, we have permanently distorted beliefs. • CLM (2000) pessimistic views come from agents assigning a larger probability to the bad (“depression”) state.
CS model consumption as: Δln(Ct) = μ(St) + εt, where St is a state (H, L) variable. St moves according to the transition matrix F, with elements Fij = Prob[St=j|St-1=i] -a Hamilton process. Note: Fll, the probability that a contraction will continue is estimated at 0.515 with a standard error of 0.264. A 90% C.I.: [.079, 0.951], which implies that contractions could have a median durations ranging from 3 months to 13 years. => even with 100 years of data, big model uncertainty persists.
Back to CS • CS allows Average Joe to learn their way out of their pessimism. • Simplify model by setting εt=0 (simpler learning problem), but assume: • gt = 1+ μ(h)/100 if St=H (=1) • = 1+ μ(l)/100 if St=L (=0) • CS also take CLM’s μ(h) and μ(l) –known by the agents- and Fll and Fhh –unknown. Average Joe applies Bayes’s theorem to learn Fij. • Average Joe uses a beta-binomial probability model for learning about Δln(Ct). A binomial likelihood is a good candidate, a beta density is the conjugate prior for a binomial likelihood. • Average Joe has independent beta priors over (Fhh,Fll) . • Average Joe counts the number of transitions from state i to state j (ntij) through date t and learns. The agent has a prior (n0ij).
Average Joe has independent beta priors over (Fhh,Fll) . • Average Joe counts the number of transitions from state i to state j (ntij) through date t and learns. The agent has a prior (n0ij). Then: • Average Joe counts the number of transitions from state i to state j (ntij) through date t and learns. The agent has a prior (n0ij). • The updating (learning) rule: • nt+1ij = ntij + 1 if st+1=j and st=i, • nt+1ij = ntij otherwise.
The date-t estimate of F is formed from the counters: • The agent makes decisions based on the values in this F matrix. The values are treated as constants when making decisions, but are random variables until convergence to RE equilibrium. • CS need a starting point, where agents are “pessimistic” (or with “shattered” beliefs, because of the depression). Think of a “worst case” F=FWC. (The WC model should be hard to reject in a training sample T0.) • Note: This is partial equilibrium, as many in the asset pricing literature. • The agent cannot affect the system by changing beliefs. Hence, there is no active learning incentive. Agents “wait.” Learn about Fhl complicated.
Estimation • CS draw 1, 000 consumption growth paths of 70 years each, assuming the true Markov chain given by F and μ(h) and μ(l). • Pessimistic Average Joe is endowed with FWC. Then, he determines asset prices using Euler’s equation and applies Bayes rule each period. • Pt(St=i,Ct) = βΣj Fij(t) (gj,t+1)-γ [Pt+1(St+1=j, gj,t+1Ct) + gj,t+1Ct]
Findings: Two anomalies are explained • (1) High market premium, but low risk aversion (from surveys) –i.e, the equity premium. • (2) Ex-post arbitrage opportunities: Euler equations hold ex-ante, but with respect to the agents’ subjective F, not the ex-post realized frequencies. • Note: Slow convergence is critical to explain anomalies.
Survival • Ex post (observed) Et[rt+1 – rf ] > ex ante Et[rt+1 – rf ] • Ex post Et[rt+1 – rf ] decreases with crash frequency • Cross-sectional implications - Low risk markets have higher ex post premium • International evidence: From Goetzmann and Jorion (1999) • U.S. market shows the maximum equity risk premium (5.48%) • 6 of 21 markets experienced no interruption from the 1920's. • 8 had a temporary closure and 7 suffered a long-term closure. • The “Non-U.S. Survived markets” returned 4.52%. while the “Non-U.S. All markets” returned 3.84%. • Survival bias is around .60% • The GJ evidence points more towards a “good draw” for the U.S.